xingyaoww

/

Qwen2.5-Coder-32B-Instruct-AWQ-128k

like 6

Is vLLM the only platform/engine that can use YARN rope scaling? I only use ooba as im an ignorant user. I cant find reference to YARN under webui on the qwen page, only vLLM. So i guess no 128k for ooba?